Interpretable Representation Learning for Wound Healing Dynamics
Prof. Bianca Dumitrascu (Columbia University)
Special DMLS Seminar
- 12:30 UZH Y35-F-32 Irchel Campus
Abstract Single-cell RNA-seq enables the study of cell states across diverse biological conditions, such as aging, drug treatments, and tissue injury. However, disentangling shared and condition-specific transcriptomic patterns remains a significant computational challenge, particularly in settings with missing data or complex experimental designs. In this talk, I will introduce Patches, a deep generative framework designed to disentangle these transcriptomic signals, allowing for robust integration, cross-condition prediction, and biologically interpretable insights. Using real and simulated scRNA-seq datasets, we demonstrate how Patches uncovers shared wound healing patterns and distinct changes in cell behavior, including age-dependent immune responses and drug-modulated extracellular matrix remodeling. Finally, I will discuss open problems towards a pipeline for synthetic regeneration which includes identifying and designing targeted therapeutic interventions to accelerate wound healing.
Bio Bianca Dumitrascu joined Columbia University in the Spring 2023 as an Assistant Professor of Statistics and Herbert and Florence Irving Assistant Professor of Cancer Data Research. Before joining Columbia, she was a Departmental Early Career Fellow at the University of Cambridge. Previously, she was a Member of the School of Mathematics at the Institute for Advanced Study. She earned her PhD in Computational Biology at Princeton University, focusing on experimental design in single-cell gene expression studies and methods for high-dimensional medical and genomic data, and completed her undergraduate studies in Mathematics at MIT.
Professor Dumitrascu leads the Computational Morphogenesis Lab where her research bridges machine learning and genetics to uncover how molecular interactions shape emergent patterns in biological systems. Inspired by the processes of regeneration and wound healing, she develops computational methods with roots in statistics, graph neural networks, domain adaptation, and active learning to advance understanding in spatial transcriptomics and cellular development. Her work sheds light on how cells coordinate and adapt in response to injury or during early developmental processes.